Diane Greene at Startup School 2013

Hi there. I've been in this auditorium once before. I think it was before you were born; it was 1989. I was working for Tandem Computers, which was one of the biggest companies in Silicon Valley. The very wonderful, irreverent founder CEO was holding an all-hands meeting, and I was in the audience as a software engineer. So it's pretty special to be here.

Today, I've done three startups. The first one was a low bandwidth streaming video company that Microsoft bought; it was the basis for their movie player and came out of a Silicon Graphics internet television project. Then I'm doing a startup right now, which I'm not going to talk about. The second startup I co-founded was VMware, and that's what I’m going to try and sort of tell you a little bit about—how we got going on VMware and how that company started.

Just by way of background, we founded the company in 1998. It's virtualization; we created sort of the virtualization industry. What it is, it's a layer of software that sits between the hardware and the operating system, kind of fakes out the operating system to think it's running on the hardware. It lets you run multiple operating systems that can be moved around on a single computer or across computers, and it works on desktops and servers. It increases efficiency, makes things easier to use, and so forth. So that was what we did. I left the company in 2008; our run rate was about $2 billion. We grew my last quarter; we had year-over-year growth of 54%, about 6,000 people in offices all over the world. It continues to be an even bigger company today.

Okay, so one thing we did when we started VMware was we had a big vision, to lessen constraints and facilitate innovation for other system builders, as well as end-to-end users of computers. We also pushed ourselves. We said it would be ubiquitous by the year 2000. We missed by a good six years or so, but we did make it ubiquitous. Then we had some students leave school on us, so we had a goal of getting them to get their Ph.D.s. Interestingly enough, one of them, after getting married and having two kids in grade school, went back to Stanford and finished his Ph.D. He's a professor in Switzerland now.

But we had this big vision; this was our founding team minus myself. I took the picture, and it has my husband, who’s a professor of computer science at Stanford, two of his graduate students at Stanford, and then my husband I met at UC Berkeley in graduate school, computer science. The other person came from there, and then one was there—that was employee number one. He was an undergraduate working on the research project, and I’m proud to say that he has gone on to found his own company, which Google recently bought.

Just for historic reasons, for those of you that are local, this was our first office over the Cheese Board in Town and Country Shopping Center. I know people are paying upwards of $5 or $6 a square foot today; we were a little luckier; it was $1.75 when we got our first office. One little bit of advice we did—we made sure we had desks to sit at, but we didn’t really spend much fun fixing up the office. I brought in an unbelievable office manager; she was more like the Chief Operating Officer when we were about ten people, and that’s when we fixed up the office.

One other thing worth mentioning about us and our founding: once we had decided to found the company, I became pregnant with our second child. I wasn’t planning to take the company all the way; I said, “Look, I’ll get it off the ground. I know how to do a startup, but I’m just having my second kid.” But for all the family people in the audience, it worked just great. I actually would have left if it wasn’t working. My daughter was jumping for joy when I told her I was about to do a third startup; she’s in high school now, so it couldn’t have been too bad for her. You just have to include them. Like, our son’s fifth birthday, we moved from this office to another, and we rented a pickup truck, which was of course his dream—to get to drive and sit in our lap and drive a pickup truck. So, we really integrated the family.

There were five of us funding-wise. What we did was self-funded initially. We said immediate family could put in money, but no friends. We weren’t totally, I mean, it took us many hours the first time— I think about four hours—to boot Windows on Linux the first time, and I was a little nervous that we were going to get the performance down. Once we convinced ourselves that this was really going to work, we went out for angel money, and we very carefully chose our angels to be people that were very technically sophisticated, that would understand what we were doing. They were computer scientists—Andy Bechtolsheim, John Hennessy, and David Sheridan. Because they understood it so well, it only took a phone call, really, you know, which is how a lot of angel investing I think happens.

Once we got a little further along, once we got into the market, we decided we needed more runway; we would take some more money and do a real outside round. We decided we were selling a product at that point—a desktop product. So we thought, “Well, let’s get some partners to invest in us because they’ll be strategic.” Since we had—we were going to start out on the workstation; we were going to go into the server once we got into the server—that would be a really difficult market to go into. So let’s get the hardware vendors to invest in us.

We looked—we decided that since we had—IBM, Dell, or HP. We thought IBM and HP could build software, so they wouldn’t be good partners in the long term perhaps. But Dell, we didn’t expect—today they do software, but back then there was no chance they were going to do software. So we went to Michael Dell and convinced him to invest in us and take the lead on a round. We were going through that round; this was in 2000, at sort of the height of the bubble. Some VCs found out we were doing a round, some really great VCs, and they said, “What are you doing? You should take our money; we’ll give you—you know.” We said, “Well, we decided we would keep the valuation where it was.” They were willing to take it higher, but we wanted to go public; we didn’t want to ever risk a down round. We believed the bubble was going to burst, so we held our valuation, and I kept—I didn’t tell Dell that we were looking at doing VCs; I just kept that term sheet going.

We moved forward with the VCs, and then in April of 2000, the bubble burst. The VCs called up and said, “We’re going to have to have your valuation; look what’s happened; everybody’s half their valuation now.” We were able to say, “No, we don’t need to do that,” and we closed the round the next day with Dell. It’s just sort of keep your options open.

What was I going to say about the funding? Oh, I’ve got to say that we were—people were running Windows on Linux with our product. Michael Dell, who was our lead investor, was quoted in CNET as saying, “I invest in lots of Linux companies; none of them are ever going to make a profit. I just want to help Linux.” When you’re an entrepreneur, you just don’t like seeing that in the press about your company. Fortunately, he was wrong, the downside of non-professional investors. But he was very helpful to us.

So we had money, we had a place to live, and we were building our product. There was a lot of very deep technology that we had to do, but we had to figure out what were going to be our first—when you have a giant vision, the thing you want to do is find a very doable first milestone that creates value for somebody. It could be another company, ideally a customer, but it doesn’t have to be a customer. So what smaller thing could we do rather than our big vision? We thought, “Well, we had this x86 virtualization; it would be a really great way for someone that was running Linux.” There were a lot of developers using Linux back then; it was just starting to rise. Linux had not quite yet been on the cover of Time. So we would build a desktop product that would let you run Windows on Linux, and that way we could also let Linux run on Windows. We figured that way we could use the Windows device drivers and not have to build every device driver that a PC needs. We wouldn't have to go into the server market, and we also had a rule that we couldn't have a dependency on anybody. We didn’t want to require Intel’s x86 architecture to make any concessions for us because that would give them a lot of power over us.

We couldn't require Microsoft to make any concessions for our software to run because they would obviously have complete control over us. So we set out to build this desktop product that would run Windows on Linux or Linux on Windows but require zero change for any operating system or chipset vendor. It had to have adequate performance so someone would want to use it, and we sort of set that as our first milestone. We got it working and said, “Okay, how do we take it to market?” We’ll take it to market as a Linux tool so that when you want to run, you know, back then, everybody that worked in a company had to run Microsoft Outlook, Microsoft Office, but they wanted to live in a Linux world. So this was the way to do it without having to have more than one computer or having to dual boot.

The Linux community would be a very technical, friendly community for us. So, that’s what we did. We launched it; everybody thought we were a Linux tool. We had three Linux companies come to us and try and buy us. We didn’t know how much we could charge for that; we got to price your first product. We said, “Well, we’ll just put the beta up there; we’ll make it fully functional, free. Anybody can have it over the internet, but they have to—it’s only a 30-day trial. They can renew it every 30 days if they want to go through the hassle because we really weren’t sure anybody would pay for it.”

Then we said, “Well, if you use it commercially, if you’re using it in a company, you should pay $299, and if you don’t want to have to renew every 30 days and you’re just a hobbyist, you should pay $99.” So it was this total trust system. We didn’t know how it would work, but we just put it up on the internet like that. We had to figure out how to spread the word when we launched this thing.

We decided we would go to a conference that launched startups—Demo. That was a funny anecdote; somebody told us we needed to hire a marketing person to help us get into demos. We hired the person; they brought the person in; they recommended to talk to them. At a demo, you can either be on stage, or you can just have a little table. Everybody gets a table, but you have to audition to get on stage. She said you’ll never get on stage, and actually, we couldn’t even recruit a PR firm at the time—everybody thought this was such a bad idea, because this was when Webvan was out there, Toys.com was out there, Pets.com was out there.

So anyhow, we decided we would launch our product at Demo. We fired that marketing person; we didn’t hire her. We did it ourselves; we auditioned. We had a funny audition because we brought in two computers; you didn’t really have laptops, and one was a backup in case the thing didn’t work. I’m sitting there with the two founders, and the women were auditioning for us. All of a sudden, the founder just froze—the keyboard was up in the air, and I’m like, “I wonder what that was about.” I kept talking away about all the ways you can use this, and then it starts working; he gets his keyboard up, and it’s working. What had happened was it had crashed, and he wanted an excuse to switch computers, so he pretended like he had dropped his keyboard.

But we got on stage, and we actually got a— it was really a wonderful experience. We got a standing ovation because we did the Windows blue screen of death! But we hadn’t really crashed the computer; we had just crashed the virtual machine, and we switched into a new window, and that got us a standing ovation.

So we did Demo, we did Slashdot, which was where all the Linux developers exchanged information. We got Slashdotted, and then we decided not to spend any money on marketing but to get public relations—to get press because we thought it was interesting enough, and it worked. We got a lot of great coverage, including in the Wall Street Journal. Actually, what happened was when we launched our beta, we were very frugal, and we didn’t have enough bandwidth. We launched it on a Sunday; we had announced we would launch it on Monday. So many people came to our site in the middle of the night to download our product that we ran out of bandwidth, and our ISP wouldn’t give us any more bandwidth automatically.

A kid at Cornell, a sophomore at Cornell, took our site, hosted it at Cornell, and posted a sign: “Don’t sue me, VMware. I’m keeping track of everybody that comes to your site, but your site is down.” He hosted us until we could get back up and running again, and he came and interned with us after that. That was great.

So anyhow, we got all that done. We also filed patents. I’ll just say a bit about patents—we started working on patents from day one. When you sit in between Microsoft and Intel, you sort of want to try and protect yourself with IP. I have to say, for a little startup, it’s pretty hopeless to protect yourself against a monopoly because they can afford giant legal bills and you can’t. But we did get our patents filed.

I can’t say, even though they were pretty rock-solid patents, they never really helped us that much. Microsoft actually bought someone that had copied us, and we weren’t going to ever sue them because we knew we were going to beat them. But then, when we found out Microsoft was trying to buy them, we decided to sue them because certainly Microsoft wouldn’t buy a company that was violating our rock-solid patent. But we were wrong.

It still took—it actually ended up slowing Microsoft down because that company’s code wasn’t very good, and Microsoft had issues with it. When I went out to look for a board of directors, since we had sort of self-funded and we’d had the angels, we just had our own internal board. Then we finally said, “Well, we should get some external people.”

Again, we did have a lot of trepidation about wedging ourselves in between Intel and Microsoft. I kind of went out and looked for who I thought was the best able to help us figure out how to stand up to those guys, and Larry Sonsini was the managing partner of Wilson Sonsini—pretty powerful attorney, business-minded attorney. So I decided he needed to be our board member. I’d never met him, and this is sort of a philosophy I have.

If you have a good reason for someone to help you that goes beyond the fact that they’re rich and powerful, you can generally convince them. I did—that he had never met me. I told the lawyer at Wilson Sonsini that we would move to their law firm if I could have a meeting with Larry Sonsini. I got my meeting; I sent him a bunch of materials over before I arrived about what we were doing, and I went into his office. I wasn’t sure I wanted him on my board, but he smiled at me and asked how he could help, and I thought, “Yes, I want him on my board.”

I said, “I’d really like you to join my board.” He laughed and said he was busy getting off boards. I said, “Well, let’s talk about the company.” So we started talking, and we started talking about—we had had quite a history with Intel around working with, shall we say, they were interested in our IP, and we were not about to give it to them.

There were some discussions going on, so I was talking to him about this and some things that were already going on with Microsoft, and he was really interested. It was late Friday afternoon; I left. He said he’d let me know, and I came in Monday morning at 7:00 a.m. There on my voicemail from six in the morning was a voicemail from Larry Sonsini saying he would join our board. He was just an invaluable board member because he had seen every deal in the world and knew how to negotiate these things.

From there, I added some other board members, but I think a board can be really helpful to you if you get the right people. Let’s see. We basically launched our product. When we started selling it, we just sold it over the internet. We supported it via email and so on. We were very frugal, and we were cash neutral. We were international; we were all over the world from day one. It was pretty exciting back in 1999.

Now, I’m going to just talk a little bit about the enterprise market because it’s sort of daunting to go into the enterprise market. In fact, I remember one of our top engineers left; he said, “You’re not going to be able to get into the enterprise market; this company is going to fail.” I remember talking to one of the better-known tech bankers, and he made fun of us. He said, “You can’t make it in the enterprise; you can do these things, but you really got to build up your credibility, and you really got to play the win shifts to get there.”

We decided we should go to market. We couldn’t possibly build out an enterprise sales force, and we should go to market with the hardware vendors. We found a way that we could help the hardware vendors make money. Basically, IBM at the time—now IBM had invented virtualization back in the 60s. It had completely died out because people thought it was something for big mainframes. But they got what we were doing, and IBM had some huge x86 servers—PC servers—that they were having trouble selling because nobody had a workload that was big enough to need such a big server.

So we were talking to them; we were talking to HP. We were trying to convince anybody that would listen to us. Just to give you a point about how little people listened to us, we were trying to talk to all the analysts like Gartner and Forrester. One day, we had one marketing guy run into my office; he goes, “We finally got a meeting with Forrester! It’s at nine o’clock tomorrow morning in Boston.” I, you know, we were in Palo Alto, and I’m like, “Okay.” So we get on the red-eye, and we go out to Boston, and we come into Forrester. We’re so excited to have the audience of an analyst, and the person they sent in was covering the auto industry.

It was like I nearly jumped at him, but I didn’t. I proceeded to tell him what virtualization was. But anyhow, finally, we were trying to get these hardware vendors to work for us, and what happened was an IBM fellow came knocking at our door. We hadn’t launched our server product yet, and he says, “I’ve been looking at your desktop product. I think it would be great on one of our servers.” A very technical guy—we’d only been talking to the marketing people—so we said, “Oh, come into our labs,” and we showed them. We had been building it from day one, and it was working.

So we got our partnership with IBM going, and then we created this sort of artificial VMware preferred hardware vendor program and set a deadline. They put a bunch of requirements on it, and all the other hardware vendors signed up by that deadline. I’m not sure they knew why they were signing up, but having a deadline helps in almost everything. We were able to start selling with the hardware vendors.

It’s an interesting sort of story about how we got going with IBM. Startups are one lucky break after another, and the game is to put yourself in position to take advantage of all these lucky breaks. What happened was IBM was going to resell our product. But we used Linux as a way to get going; it was kind of boot, and then Linux got out of the way, and then our software took over. You didn’t use it anymore, and IBM’s lawyers discovered that we had GPL Linux software in our distribution, and they killed the whole deal because they were afraid it would pollute everything, and all of IBM’s software would suddenly have to be open-sourced. They didn’t want to risk that.

We were like, “Oh no, how are we ever going to get anyone to help us sell our product?” Then we had this idea. We said, “Well, what if your resellers sold us? Would that create this legal problem?” They said, “No, our resellers can sell you without it causing any problems.” So we took that to them, and what ended up happening was they were about to have their worldwide conference of all their top-tier resellers in the world. By the way, apologies for all of you that are in consumer and don’t care a bit about the enterprise, but it’s an interesting story.

All these resellers—these are the best-trained people selling this kind of product. They invited us to the conference; they let us keynote; they let us run training sessions. After that, our server market took off. It was kind of a lucky break to be able to do that.

The thing I wanted to close on, and I even— I couldn’t believe it. I experienced it here today. I was walking around, having a piece of pizza before I came inside, and somebody grabbed me and said, “Are you the founder of VMware? Thank you for your product. I love your product.” That hasn’t happened in a few years, so thank you. But anyhow, that’s really what it’s about: bringing something to market that people absolutely love and makes their lives better.

This was my favorite email we ever received: “Congratulations for that. It made it an amazing product. I can’t believe what I see; I think it’s the best discovery after landing on the moon. The only thing I can’t believe is—why is it beta? Other software— we knew they meant Microsoft isn’t that stable in a final version.” Somebody else said, “I can’t believe it; your brains must be bigger than Volkswagens.” Then this incredible article that the Wall Street Journal did about bringing freedom of choice to the desktop.

It’s things like that. It was just such a wonderful, fun experience to bring something so useful to people to market. Well, I thought there would be questions, but I guess it's hard with this many people. I love getting questions. I was just going to say good luck to all of you that might be doing a startup. I’m doing another one too, and each one is different and challenging. A lot of people kind of pay attention; they think, “Oh, what are you doing?” I’m like, “I’m just another slob trying to do a startup. You know, we’ll see.”

But good luck to you, and thank you very much!